[Feature] Build and run C++ Python tests

.github/workflows/unit_test.yaml

@@ -0,0 +1,55 @@
+on:
+  workflow_dispatch:
+  pull_request:
+  push:
+
+jobs:
+  run_unit_test:
+    name: Run unit tests on ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - {os: ubuntu-latest, arch: x86_64}
+          - {os: ubuntu-24.04-arm, arch: aarch64}
+          - {os: windows-latest, arch: AMD64}
+          - {os: macos-14, arch: arm64}
+          - {os: macos-13, arch: x86_64,}
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          submodules: recursive
+
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v5
+        with:
+          python-version: 3.11
+
+      - name: Build xgrammar from source
+        run: |
+          echo "set(XGRAMMAR_BUILD_CXX_TESTS ON)" >> cmake/config.cmake
+          python -m pip install --upgrade pip
+          pip install -e .
+
+      - name: Run C++ tests
+        run: |
+          cd build && ctest && cd ..
+
+      - name: Run Python tests
+        env:
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
+        if: env.HF_TOKEN != ''
+        run: |
+          pip install pytest
+          pytest
+
+      - name: Run Python tests without HF_TOKEN
+        env:
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
+        if: env.HF_TOKEN == ''
+        run: |
+          pip install pytest
+          pytest -m "not hf_token_required"

cpp/support/thread_pool.h

@@ -12,6 +12,7 @@
 #include <mutex>
 #include <queue>
 #include <thread>
+#include <type_traits>
 #include <vector>
 
 #include "logging.h"
@@ -67,9 +68,8 @@ class ThreadPool {
    * \note Tasks are executed in FIFO order but may complete in any order.
    */
   template <class F, class... Args>
-  auto Submit(F&& f, Args&&... args)
-      -> std::shared_future<typename std::result_of<F(Args...)>::type> {
-    using return_type = typename std::result_of<F(Args...)>::type;
+  auto Submit(F&& f, Args&&... args) -> std::shared_future<std::invoke_result_t<F, Args...>> {
+    using return_type = std::invoke_result_t<F, Args...>;
 
     // Package the task with its arguments into a shared pointer
     auto task = std::make_shared<std::packaged_task<return_type()>>(

tests/cpp/test_thread_safe_cache.cc

@@ -34,7 +34,9 @@ using namespace std::chrono_literals;
 TEST(XGrammarParallelTest, CacheEfficiency) {
   auto cache = ThreadSafeCache<std::string, MockGrammar>{[](const std::string&) {
     std::this_thread::sleep_for(1s);  // simulate a slow operation
-    return MockGrammar{.uuid = counter++, .padding = {}};
+    MockGrammar g{};
+    g.uuid = counter++;
+    return g;
   }};
   auto futures = std::vector<std::future<std::size_t>>{};